GCC & LLVM Developers May Begin Collaborating

[phoronix]

Phoronix: GCC & LLVM Developers May Begin Collaborating

As an interesting turn of events after Richard Stallman called LLVM a "terrible setback" and the discussion that ensued, it turns out that the GCC and LLVM/Clang developers might start to better collaborate under some sort of open-source compiler initiative...

http://www.phoronix.com/vr.php?view=MTU5Njk

[c117152]

BSD being interested in collaboration is no surprise. It's the GNU guys not shoeing them off the door that is the real shocker.

I keep envisioning Saint IGNUcius bursting into an LLVM\GCC boardroom-meeting wielding a plastic lightsaber and wearing a modified Don Quixote's armour while shouting "Freedom!" while smacking people about...

[jakubo]

i actually thought that variety was a blessing in the compiler world, so one can better check, why one is working while the other isn't...

[Annabel]

"The Free Compiler Initiative" would be better

[jsdyson]

i actually thought that variety was a blessing in the compiler world, so one can better check, why one is working while the other isn't...

When doing commercial development (from the user standpoint, not the compiler developer), I started (about 2yrs ago) doing compiles on all of LLVM, GCC and ICC to guarantee compatibility. Unless needing to use special features of each of the compilers, I tend to make sure that my commercial code works based upon slightly different opinions about the spec. It is quite nice that there are several either free or free-to-get compilers :-).

John Dyson

[erendorn]

i actually thought that variety was a blessing in the compiler world, so one can better check, why one is working while the other isn't...

Indeed, but for some initiatives like non-standard extensions, or interfaces, or options, you cannot check on several compilers unless they try to implement the same thing.

[Bjotn]

That he is surprised that it didn't turn into a flamewar says something...

[FLHerne]

That he is surprised that it didn't turn into a flamewar says something...

He's suffering from overexposure to his forums, obviously.

[blackiwid]

For me all of that are just some talk.
RMS basicly said release everything with gpl3+ lisense, thats normal, I agree to that.

Who cares about freedom should do so, but that was the case before Richard told that anyway.
Did he fear that some people belived that when they release their stuff under bsd that its free then?

And to this point its clear, the gcc devs can take all the code and include it and make out of this peaces gpl3 code.

The other way around thats not working. So I see no problem for gcc, just everybody who cares about freedom release his stuff under gpl lisence who not cares use bsd lisenses.

K that rms again says how bad apple and this stuff is ist ok as a reminder but has no direct effect I think?

So this working together attempt make no sence because gpl devs can use both code anyway it would only help the unfree stuff?

Why would somebody from the gcc team want to do that?

[the303]

For me all of that are just some talk.
RMS basicly said release everything with gpl3+ lisense, thats normal, I agree to that.

Who cares about freedom should do so, but that was the case before Richard told that anyway.
Did he fear that some people belived that when they release their stuff under bsd that its free then?

And to this point its clear, the gcc devs can take all the code and include it and make out of this peaces gpl3 code.

The other way around thats not working. So I see no problem for gcc, just everybody who cares about freedom release his stuff under gpl lisence who not cares use bsd lisenses.

K that rms again says how bad apple and this stuff is ist ok as a reminder but has no direct effect I think?

So this working together attempt make no sence because gpl devs can use both code anyway it would only help the unfree stuff?

Why would somebody from the gcc team want to do that?

Exactly my thought.

RMS's view on BSD LLVM is clear and not positive. I don't see why GCC devs would bother with LLVM.